1. Field of the Invention
The present invention generally relates to wireless communication systems and, more particularly, to the transmission of acknowledgement information in an uplink of a communication system.
2. Description of the Art
A communication system includes a DownLink (DL) that conveys transmission signals from a Base Station (BS), or NodeB, to User Equipments (UEs), and includes an UpLink (UL) that conveys transmission signals from UEs to the NodeB. A UE, which is also commonly referred to as a terminal or a mobile station, may be fixed or mobile and may be, for example, a wireless device, a cellular phone, or a personal computer device. A NodeB is generally a fixed station and may also be referred to as an access point or some other equivalent terminology.
The UL conveys transmissions of data signals carrying information content, transmissions of control signals providing control information associated with the transmission of data signals in the DL, and transmissions of Reference Signals (RSs), which are commonly referred to as pilot signals. The DL also conveys transmissions of data signals, control signals, and RSs.
UL data signals are conveyed through a Physical Uplink Shared CHannel (PUSCH) and DL data signals are conveyed through a Physical Downlink Shared CHannel (PDSCH).
In the absence of a PUSCH transmission, a UE conveys UL Control Information (UCI) through a Physical Uplink Control CHannel (PUCCH). However, when it has a PUSCH transmission, a UE may convey UCI together with data through the PUSCH.
DL control signals may be broadcast or may be sent in a UE-specific nature. Accordingly, UE-specific control channels can be used, among other purposes, to provide UEs with Scheduling Assignments (SAs) for PDSCH reception (DL SAs) or PUSCH transmission (UL SAs). The SAs are transmitted from the NodeB to respective UEs using DL Control Information (DCI) formats through respective Physical DL Control CHannels (PDCCHs).
The NodeB may configure a UE through higher layer signaling, such as, for example, Radio Resource Control (RRC) signaling, a PDSCH and a PUSCH Transmission Mode (TM), and other parameters relating to reception of DL signals or transmission of UL signals. The PDSCH TM or PUSCH TM is respectively associated with a DL SA or a UL SA, and defines whether a respective PDSCH or PUSCH conveys one data Transport Block (TB) or two data TBs.
PDSCH or PUSCH transmissions are either scheduled to a UE by the NodeB through higher layer signaling or through physical layer signaling (through, for example, the PDCCH) using a respective DL SA or UL SA, or correspond to non-adaptive retransmissions for a given Hybrid Automatic Repeat reQuest (HARQ) process. Scheduling by higher layer signaling is referred to as Semi-Persistent Scheduling (SPS). Scheduling by PDCCH is referred to as dynamic. A PDCCH may also be used to release SPS PDSCH. If a UE misses (i.e., fails to detect) a PDCCH, it also misses the associated PDSCH or PUSCH. This event is referred to as DTX (Discontinuous Transmission).
The UCI includes ACKnowledgment (ACK) information associated with a HARQ process (HARQ-ACK). HARQ-ACK information may consist of multiple bits corresponding to positive acknowledgments (ACKs) for TBs a UE correctly received, or corresponding to Negative ACKnowledgements (NACKs) for TBs the UE incorrectly received. When a UE does not receive a TB, it may transmit DTX (tri-state HARQ-ACK information) or it may transmit a NACK that represents both the absence and the incorrect reception of a TB (in a combined NACK/DTX state).
In Time Division Duplex (TDD) systems, DL and UL transmissions occur in different Transmission Time Intervals (TTIs) which are referred to as subframes. For example, in a frame comprising of 10 subframes, some subframes may be used for DL transmissions and other subframes may be used for UL transmissions.
FIG. 1 is a diagram illustrating a frame structure for a TDD system.
Referring to FIG. 1, a 10 millisecond (ms) frame consists of two identical half-frames. Each 5 ms half-frame 110 is divided into eight slots 120 and three special fields. The three special fields include a DL ParT Symbol (DwPTS) 130, a Guard Period (GP) 140, and an UL ParT Symbol (UpPTS) 150. The length of DwPTS+GP+UpPTS is equal to one subframe (1 ms) 160. The DwPTS may be used for the transmission of synchronization signals from the NodeB, while the UpPTS may be used for the transmission random access signals from UEs. The GP facilitates the transition between DL and UL transmissions by absorbing transient interference.
The number of DL and UL subframes per frame can be different, and multiple DL subframes may be associated with a single UL subframe. In associating multiple DL subframes with a single UL subframe, a number OQ-ACK of HARQ-ACK information bits generated in response to PDSCH receptions (data TBs) in multiple DL subframes needs to be transmitted in a single UL subframe. This number of DL subframes Nbundle is referred to as bundling window.
A first method in which a UE conveys HARQ-ACK information in a single UL subframe, in response to PDSCH receptions in multiple DL subframes, involves HARQ-ACK bundling. In HARQ-ACK bundling the UE transmits an ACK only if it correctly receives all data TBs and transmits a NACK in all other cases. Therefore, HARQ-ACK bundling results in unnecessary retransmissions and reduced DL throughput, since the NACK is transmitted even when a UE incorrectly receives only one data TB and correctly receives all other data TBs.
Another method in which a UE conveys up to 4 bits of HARQ-ACK information in a single UL subframe, in response to receptions of data TBs in multiple DL subframes, involves HARQ-ACK multiplexing, which is based on PUCCH resource selection.
An additional method in which a UE conveys multiple HARQ-ACK information bits in a single UL subframe, in response to receptions of multiple data TBs in multiple DL subframes, involves joint coding of the HARQ-ACK information bits using, for example, a block code such as a Reed-Mueller (RM) code.
If a PDSCH conveys one TB, the respective HARQ-ACK information consists of one bit which is encoded as a binary ‘1’ (ACK value) if the TB is correctly received, and is encoded as a binary ‘0’ (NACK value) if the TB is incorrectly received. If a PDSCH conveys two TBs, in accordance with the Single-User Multiple Input Multiple Output (SU-MIMO) transmission method with a rank higher than one, the HARQ-ACK information consists of two bits [o0ACK o1ACK], with o0ACK for the first TB and o1ACK for the second TB. However, if a UE applies bundling in the spatial-domain for the 2 HARQ-ACK bits associated with the reception of the 2 TBs when a SU-MIMO PDSCH has a rank larger than one, the UE feedback consists of only one HARQ-ACK bit that has the binary value 0 (NACK value) when at least one TB is incorrectly received, or the binary value 1 (ACK value) when both TBs are correctly received. As the PDSCH TM determines a number of conveyed TBs (one or two), it also determines a respective number of HARQ-ACK bits (if spatial-domain bundling is not applied).
FIG. 2 is a diagram illustrating a PUCCH structure in one subframe slot for transmitting multiple HARQ-ACK information bits using a Discrete Fourier Transform Spread Orthogonal Frequency Division Multiplexing (DFT-S-OFDM) transmission method.
Referring to FIG. 2, after encoding and modulation, using for example, a RM block code and Quadrature Phase Shift Keying (QPSK), respectively, a set of same HARQ-ACK bits 210 is multiplied at multiplier 220 with elements of an Orthogonal Covering Code (OCC) 230, and is subsequently DFT precoded at DFT precoder 240. For example, for 5 symbols per slot carrying HARQ-ACK bits, the OCC has length of 5 {OCC(0), OCC(1), OCC(2), OCC(3), OCC(4)}, and can be {1, 1, 1, 1, 1}, {1, exp(j2π/5), exp(j4π/5), exp(j6π/5), exp(j8π/5)}, {1, exp(j4π/5), exp(j8π/5), exp(j2π/5), exp(j6π/5)}, {1, exp(j6π/5), exp(j2π/5), exp(j8π/5), exp(j4π/5)}, or {1, exp(j8π/5), exp(j6π/5), exp(j4π/5), exp(j2π/5)}. The output of the DFT precoder 240 is passed through an Inverse Fast Fourier Transform (IFFT) unit 250 and it is then mapped to a DFT-S-OFDM symbol 260. Since the previous operations are linear, their relative order may be inter-changed. Because a PUCCH transmission is assumed to be in one PRB, which consists of NscRB=12 REs, there are 24 encoded HARQ-ACK bits transmitted in each slot (12 HARQ-ACK QPSK symbols) and a (32,OHARQ-ACK) RM code is punctured into a (24, OHARQ-ACK) RM code. The same or different HARQ-ACK bits may be transmitted in the second slot of a subframe. In addition to HARQ-ACK signals, RS are transmitted in each slot to enable coherent demodulation of HARQ-ACK signals. Each RS is constructed from a length-12 Zadoff-Chu (ZC) sequence 270, which is passed through an IFFT unit 280 and mapped to another DFT-S-OFDM symbol 290.
The PUCCH structure in FIG. 2 can support reliable reception for only a limited number of HARQ-ACK information bits, which is also referred to as a HARQ-ACK payload, without incurring a large coding rate as it can only support 24 encoded HARQ-ACK bits. The use of a dual RM code can allow for support of larger HARQ-ACK payloads. For example, a single RM code can be used for HARQ-ACK payloads up to 10 bits, and a dual RM code can be used for HARQ-ACK payloads between 11 and 20 bits. With a dual RM code, the mapping to successive elements of the DFT can alternate between elements from the output of a first RM code and elements from the output of a second RM code in a sequential manner. For HARQ-ACK payloads of more than 20 bits, convolutional coding can be used.
FIG. 3 is a diagram illustrating a transmitter block diagram for transmitting HARQ-ACK information encoded using a single RM code.
Referring to FIG. 3, HARQ-ACK information bits 305 are encoded and modulated by an encoder and modulator 310, and then multiplied with an element of an OCC 325 for the respective DFT-S-OFDM symbol at multiplier 320. The output of the multiplier 320 is then DFT precoded by DFT precoder 330. After DFT precoding, sub-carrier mapping is performed by a sub-carrier mapper 340, which is under control of a controller 350. Thereafter, the IFFT is performed by an IFFT unit 360, a CP is added by at CP inserter 370, and the signal is filtered for time windowing by filter 380, thereby generating a transmitted signal 390. Additional transmitter circuitry, such as, for example, a digital-to-analog converter, analog filters, amplifiers, and transmitter antennas may also be included in the transmitter block diagram of FIG. 3.
FIG. 4 is a diagram illustrating a receiver block diagram for receiving HARQ-ACK information encoded using a single RM code.
Referring to FIG. 4, after receiving a Radio-Frequency (RF) analog signal and converting it to a digital signal 410, the digital signal 410 is filtered for time windowing at a filter 420, and a CP is removed at a CP remover 430. Subsequently, the NodeB receiver applies an FFT at an FFT unit 440, performs sub-carrier demapping at a sub-carrier demapper 450, which is under the control of a controller 455, and applies an Inverse DFT (IDFT) at an IDFT unit 460. The output of the IDFT unit 460 is then multiplied with an OCC element 475 for the respective DFT-S-OFDM symbol a multiplier 470. An adder 480 sums the outputs for the DFT-S-OFDM symbols conveying HARQ-ACK signals over each slot, and a demodulator and decoder 490 demodulates and decodes the summed HARQ-ACK signals over both subframe slots to obtain HARQ-ACK information bits 495. Well known receiver functionalities such as, for example, channel estimation, demodulation, and decoding may also be included in the receiver block diagram of FIG. 4.
FIG. 5 is a diagram illustrating a transmitter block diagram for transmitting HARQ-ACK information encoded using a dual RM code.
Referring to FIG. 5, the payload of OHARQ-ACK HARQ-ACK bits 505 is first segmented into two parts of OHARQ ACK1=┌OHARQ ACK/2┐ bits and OHARQ-ACK2=└OHARQ ACK/2┘ bits at segmentation block 510. The segmented parts are subsequently individually encoded with a (32, OHARQ-ACK1) RM code and a (32, OHARQ-ACK2), respectively, and each of the 32 coded bits are then punctured to 24 coded bits which are then QPSK modulated to obtain 12 QPSK coded symbols, at coding and modulation blocks 520 and 525, respectively. The first 6 for each of the 12 QPSK coded symbols are combined, for example, by interlacing, at a block 530 and are then multiplied with an element of the OCC 545 for the respective DFT-S-OFDM symbol at a multiplier 540 for transmission in a first slot of a subframe. The same processing applies to the last 6 of the 12 QPSK coded symbols, which are transmitted in a second slot of the subframe. After DFT precoding at a DFT precoder 550, the REs of the assigned PUCCH PRB are selected at a sub-carrier mapper 565, which is under the control of a controller 560. The IFFT is performed at an IFFT block 570 and finally the CP and filtering are applied to a transmitted signal 580. Additional transmitter circuitry, such as, for example, a digital-to-analog converter, analog filters, amplifiers, and transmitter antennas may be included in the transmitter block diagram of FIG. 5.
FIG. 6 is a diagram illustrating a receiver block diagram for receiving HARQ-ACK information encoded using a dual RM code.
After an antenna receives the RF analog signal and after further processing units (such as filters, amplifiers, frequency down-converters, and analog-to-digital converters), a digital signal 610 is filtered and the CP is removed. Subsequently, the NodeB receiver applies an FFT at an FFT block 620, selects REs used by the UE transmitter at a sub-carrier demapper 630, which is under the control of a controller 635. The NodeB receiver applies an IDFT at an IDFT block 640, multiplies with an OCC element 655 for the respective DFT-S-OFDM symbol at a multiplier 650, sums the outputs for the DFT-S-OFDM symbols over each slot at a summing block 660, collects the QPSK symbols from both subframe slots at a collection block 670, splits (de-interlaces) the 24 QPSK symbols in the original pairs of 12 QPSK symbols in a split block 675, and demodulates and decodes each of the two pairs of 12 QPSK symbols at demodulation and decoding blocks 680 and 685, respectively, to obtain transmitted HARQ-ACK bits 690. Well known receiver functionalities, such as, for example, channel estimation, demodulation, and decoding, may also be included in the receiver block diagram of FIG. 6.
Using the maximum HARQ-ACK payload in a PUCCH does not create additional resource overhead. A UE may transmit a NACK or a DTX (in case of tri-state HARQ-ACK information) for the TBs it did not receive. However, the NodeB already knows the DL cells with no DL SA or PDSCH transmission to the UE, and can use the knowledge that the UE transmits a NACK for each of those DL cells (a-priori information) to improve the HARQ-ACK reception reliability. This is possible because a linear block code and QPSK are assumed to be used for the encoding and modulation of the HARQ-ACK bits, respectively, and the NodeB can consider, as candidate HARQ-ACK codewords, only those having NACK (binary ‘0’) at the predetermined locations corresponding to cells without DL SA transmissions to the UE. Due to the implementation of the decoding process, the use of the a-priori information would be impractical or impossible if a convolutional code or a turbo code was used for encoding the HARQ-ACK information bits. Therefore, the coding rate for the transmission of HARQ-ACK information in a PUCCH depends on the number of HARQ-ACK information bits the NodeB does not know in advance.
For HARQ-ACK transmission in a PUSCH, a UE determines a respective number of coded symbols Q′ as shown in Equation (1) below.
                              Q          ′                =                  min          (                                    ⌈                                                                    O                                          HARQ                      -                      ACK                                                        ·                                      β                    offset                    PUSCH                                                                                        Q                    m                                    ·                  R                                            ⌉                        ,                          4              ·                              M                sc                PUSCH                                              )                                    (        1        )            
In Equation (1), βoffsetPUSCH is informed to the UE through higher layer signaling, Qm is the number of data modulation bits (Qm=2, 4, 6 for QPSK, QAM16, QAM64, respectively), R is the data code rate of the initial PUSCH transmission for the same TB, MscPUSCH is the PUSCH transmission BW in the current sub-frame, and ┌ ┐ is the “ceiling” function which rounds a number to its next integer. The code rate R is defined as
  R  =            (                        ∑                      r            =            0                                              C              CB                        -            1                          ⁢                  K          r                    )              (                        Q          m                ·                  M          sc                      PUSCH            -            initial                          ·                  N          symb                      PUSCH            -            initial                              )      where CCB is the total number of code blocks and Kr is the number of bits for code block number r. The maximum number of HARQ-ACK REs is limited to the REs of 4 DFT-S-OFDM symbols (4·MscPUSCH). The value of Qm·R determines the Spectral Efficiency (SE) of the data transmission in the PUSCH and, given MscPUSCH, it can be directly derived from the Modulation and Coding Scheme (MCS) used for the data transmission.
In TDD systems, as a UE needs to send HARQ-ACK information corresponding to potential TB receptions over multiple DL subframes in a bundling window, a DL Assignment Index (DAI) Information Element (IE), VDAIDL, is included in the DL SAs to inform the UE of the number of DL SAs transmitted to it by the NodeB. Since the NodeB cannot predict whether there will be a DL SA for a UE in future DL subframes, the VDAIDL is a relative counter which is incremented in each DL SA transmitted to the UE and starts from the beginning after the last DL subframe in the bundling window. If the UE fails to detect the last DL SA, it cannot become aware of this event while if the UE fails to detect a DL SA other than the last one, it can become aware of this event if it receives another DL SA in a subsequent DL subframe of the same bundling window.
FIG. 7 is a diagram illustrating a setting for a DL DAI IE over 4 DL subframes of a bundling window.
Referring to FIG. 7, in a DL subframe 0 710, the NodeB transmits a DL SA to a UE and sets the DL DAI IE value to VDAIDL=0. In a DL subframe 1 720, the NodeB transmits a DL SA to the UE and sets the DL DAI IE value to VDAIDL=1. In a DL subframe 2 730, the NodeB does not transmit a DL SA to the UE and does not increment the DL DAI IE value. In a DL subframe 3 740, the NodeB transmits a DL SA to the UE and sets the DL DAI IE value to VDAIDL=2.
If a UE has data transmission in a UL subframe where it is expected to also transmit HARQ-ACK information, then both data and HARQ-ACK may be transmitted in a PUSCH. In order to avoid error cases where the UE has missed the last DL SA and ensure the same understanding between the NodeB and the UE of the number of HARQ-ACK bits in the PUSCH, a DAI IE is also included in the UL SA (UL DAI IE) scheduling the PUSCH to indicate the number of HARQ-ACK bits the UE should include. For the setup in FIG. 7 where Nbundle=4, the UL DAI IE can be represented by 2 bits with respective values of VDAIUL=0 or 4, 1, 2, 3. If the UE receives a DL SA in the bundling window, then the UL DAI IE bits of “00” map to a UL DAI IE value of VDAIUL=4 instead of VDAIUL=0.
In order to support high data rates in a communication system, Carrier Aggregation (CA) of multiple cells is considered to provide higher operating BandWidths (BWs). For example, to support communication over 60 MHz, CA of three 20 MHz cells can be used.
FIG. 8 is a diagram illustrating the principle of CA.
Referring to FIG. 8, an operating DL BW of 60 MHz 810 is constructed by the aggregation of 3 cells, DL CC 1 821, DL CC 2 822, and DL CC 3 823, each having a DL BW of 20 MHz. Similarly, an operating UL BW of 60 MHz 830 is constructed by the aggregation of 3 cells, UL CC 1 841, UL CC 2 842, and UL CC 3 843, each having an UL BW of 20 MHz.
For simplicity, in FIG. 8, each cell is assumed to have a unique DL and UL pair (symmetric CA), but it is also possible for more than one DL to be mapped to a single UL and the reverse (asymmetric CA). This mapping is typically UE-specific and the NodeB can configure a set of C cells to a UE, using for example Radio Resource Control (RRC) signaling, and activate a subset of A cells (A≦C) for PDSCH reception in a subframe, using for example Medium Access Control (MAC) signaling (a UE may not monitor inactive cells for communication with the NodeB). If a PDSCH activating or deactivating configured cells is missed, then the UE and the NodeB may have a misunderstanding of the active cells. Moreover, in order to maintain the communication, one cell with a DL/UL pair needs to remain always activated and it is referred to as the primary cell. The PUCCH transmissions from a UE are assumed to be only in its primary cell (Pcell) and HARQ-ACK information is conveyed only in a single PUSCH.
FIG. 9 is a diagram illustrating the parallelization of the DL DAI design in FIG. 7 to multiple DL cells.
Referring to FIG. 9, a NodeB transmits to a UE DL SAs in 3 DL subframes in Cell 0 910 and sets the DL DAI IE values according to the number of DL SAs transmitted to the UE only for PDSCH receptions in Cell 0. In a similar manner, the NodeB transmits to the UE DL SAs in 2 DL subframes in Cell 1 920 and 2 DL subframes in Cell 2 930 and sets the DL DAI IE values according to the number of DL SAs transmitted to the UE only for PDSCH receptions in Cell 1 and Cell 2, respectively.
A fundamental condition for proper communication of the HARQ-ACK information is for a UE and a NodeB to have a same understanding of the HARQ-ACK payload. This includes the same understanding about the ordering of HARQ-ACK information bits across cells and subframes in a transmitted HARQ-ACK codeword and of the coding method used to transmit the HARQ-ACK payload (single RM or dual RM code).
The actual HARQ-ACK payload also needs to be limited as desired reliability requirements are difficult to achieve otherwise. Additionally, the required resources in a PUSCH for transmitting large HARQ-ACK payloads can become excessive and lead to unacceptable overhead or an inability to reliably receive the HARQ-ACK payload. For this reason, the HARQ-ACK payload needs to be compressed and spatial-domain bundling is considered as the first choice, possibly followed by bundling across DL subframes (time-domain bundling) or across cells (cell-domain bundling).